Limits and caps
Boost has a lot of ceilings, and most of them only become visible when you hit one. This page states every number, and the exact sentence you see when you cross it.
Per-plan ceilings
| Limit | Free | Lite | Plus | Elite | Master | Ultimate |
|---|---|---|---|---|---|---|
| Standing discount | 0% | 5% | 8% | 10% | 12% | 15% |
| Orders in flight at once | 5 | 25 | 50 | 100 | 250 | 1,000 |
| Maximum value of one order | $25 | $200 | $500 | $1,500 | $5,000 | $100,000 |
| Rows in one batch | 20 | 25 | 50 | 150 | 500 | 1,000 |
| Auto-Boost rules at once | 1 | 3 | 5 | 8 | 15 | 50 |
| Active API keys | — | — | — | 3 | 10 | 25 |
| API requests per minute, per key | — | — | — | 60 | 120 | 300 |
| Included delivery guarantee | — | — | — | 30 days | 60 days | 90 days |
| Drip-feed | — | Yes | Yes | Yes | Yes | Yes |
| Priority delivery | — | — | Yes | Yes | Yes | Yes |
| CSV export | — | Yes | Yes | Yes | Yes | Yes |
| Developer API | — | — | — | Yes | Yes | Yes |
| Order webhooks | — | — | — | — | Yes | Yes |
| Wholesale rates | — | — | — | — | Yes | Yes |
Referral Premium tracks the Free column on every one of these rows.
The API row is what your plan sells. What a key actually gets is that figure banded by your reliability score, and it can be overridden or capped for one key — The Boost API holds those numbers.
Ultimate's plan entitlement is 2,000 batch rows, but a platform ceiling of 1,000 rows per request bites first, so 1,000 is the real number. Split anything larger across separate batches.
Platform ceilings that apply to everyone
| Ceiling | Value | Where it bites |
|---|---|---|
| Maximum quantity on one order | 100,000,000 units | Checked before the service's own minimum and maximum, which are always narrower |
| Maximum charge on one order | $100,000 | Applies even on Ultimate, whose plan cap is the same figure |
| Rows in one batch request | 1,000 | Beats a higher plan entitlement |
| Drip-feed batches | 2 to 100 | The composer offers 4, 8 and 12 |
| Drip-feed interval | 1 to 1,440 minutes | The composer offers 30, 60 and 120 |
| Active webhooks | 10 | Master and Ultimate |
| Rows in one CSV export | 10,000, newest first | Lite and above |
Rate limits
Each is counted per account unless stated otherwise. Crossing one gives you a generic too-many-requests refusal, not a specific message — which is why they are listed here.
| Action | Limit |
|---|---|
| Placing an order | 30 per hour per account, and 60 per hour per IP address |
| Placing a batch (cart or mass order) | 12 per hour |
| Quoting a price | 120 per minute |
| Requesting a refill | 20 per hour |
| Cancelling an order | 20 per hour |
| Reporting a problem on an order | 5 per 10 minutes |
| Exporting orders as CSV | 6 per minute |
| Minting an API key | 10 per hour |
| Adding a webhook | 15 per hour |
| Loading the public price list | 60 per minute per IP address |
An API key is metered per key, not per account, and the figure is not fixed: your plan's requests per minute (the per-plan table above) is banded by your reliability score, an administrator can override it for one key, and a key can carry a calls-per-day ceiling on top. All of that is stated once, on The Boost API — this page does not repeat the numbers, so there is one authority for them.
The 10-per-hour minting limit and the 15-per-hour webhook limit above are ordinary account limits and do belong here: they are counted on the signed-in dashboard, not on a key.
The individual lines inside a batch are exempt from the 30-orders-per-hour rule, because the batch endpoint's own 12-per-hour limit already bounds them. A 500-row mass order therefore places 500 orders without touching your hourly order allowance.
The duplicate-order guard
Placing the same service, the same link and the same quantity twice within 60 seconds is treated as an accidental double-submit and refused:
You just placed this exact boost — check your Orders. If you meant to order it again, try once more in a moment.
Nothing is charged. If the repeat was deliberate, wait out the minute and place it again. The guard is separate from the idempotency protection that stops a single double-clicked button from charging twice; that one is instant and invisible.
Timing rules you can feel
| Behaviour | Timing |
|---|---|
| A Boost page re-reads your account | About every 45 seconds, paused while the tab is in the background |
| The delivery sweep runs | Every 90 seconds, across all live orders |
| One order is re-polled at most | Every 120 seconds — or every 60 with Priority delivery |
| An order stuck mid-placement is refunded | After 180 seconds |
| The delivery-protection sweep runs | Roughly hourly |
| One order is picked up by that sweep at most | Every 6 hours — a claim stamp, so two sweeps never act on the same order |
| One order can be automatically refilled at most | Once every 6 hours while a drop is actually measurable; otherwise once every 7 days. Either way, 30 times in its life |
| An order's default protection window | 30 days from completion, unless it carries a longer guarantee |
The exact refusals, and what to do
| Message | Limit it names | What to do |
|---|---|---|
You already have N orders in progress, which is the limit for the <plan> plan. | Live-order cap | Wait for orders to settle, or upgrade — the message names the next plan's number |
That order is $X, above the $Y per-order limit on the <plan> plan. | Per-order cap | Lower the quantity, or split it into two orders |
That's N items — the most you can order at once on the <plan> plan is X. | Batch rows | Remove lines, or split the batch |
| Quantity must be between X and Y for this service. | The service's own range | Read the hint under the Quantity field |
| Enter a valid quantity. | Zero, negative, or above 100,000,000 | Type a real number |
| Drip-feed batches must be between 2 and 100. | Drip-feed runs | Use one of the three cadence presets |
| The drip-feed interval must be between 1 and 1440 minutes. | Drip-feed interval | Same |
| You just placed this exact boost — check your Orders. | The 60-second duplicate guard | Check Orders first; wait a minute if it really is a repeat |
| You're $X short for this order. Top up and your order will still be here. | Your balance, not a plan limit | Top up from Overview |
You're at your <plan> limit of N active API keys. Revoke one to mint another. | API key ceiling | Revoke an unused key |
| You already have 10 active webhooks. Remove one first. | Webhook ceiling | Delete one |
| CSV export is a member feature. Unlocks on Lite. | The export gate | Downloads as a small file rather than showing on screen — see Tracking your orders |
| A bare "too many requests" response | One of the rate limits above | Find the action in the rate-limit table and wait out its window |
If something goes wrong
| What you see | What it means | What to do |
|---|---|---|
| A limit refusal names a number you do not recognise | Your plan changed, or lapsed, since you last ordered | Every limit is read live from your current plan. Check your plan on the account page |
| You are refused with no message at all | A rate limit — those do not carry a specific sentence | Match the action against the rate-limit table |
| An order over your cap was accepted last month | Your plan was higher then. Caps are evaluated at order time, never retroactively | Nothing to do; the older order is unaffected |
| Two different limits seem to contradict | The tighter of a plan limit and a platform ceiling always wins | Ultimate's batch rows are the clearest case: 2,000 on the plan, 1,000 in practice |
Related
- Placing an order — where most of these are hit
- Cart and mass orders — batch rows and batch rate limits
- Pricing and discounts — the discount and margin rules behind the price caps
- The Boost API — key counts, the banded per-key rate, every refusal the gate returns and what gets a key blocked